Matching Markets: Structure, Dynamics and Design

Matching markets such as labor markets, marriage markets, and school admissions allow agents to match with each other for mutual benefit. The theory of bilateral matching has led to successful marketplace designs for high school and college admissions in various cities/countries, and the National Residency Matching Program, among other markets. However, this theory and market designs resulting from it ignore operational issues that are of central importance in many marketplaces, especially in modern, technologically intermediated markets for labor, dating, accommodation and other purposes. First, participants in these marketplaces typically do not know their preferences beforehand. Second, cancellation of matches may occur due to one or both agents involved backing out, for instance when students do not report for classes after taking admission to a school/college. Third, the platform may not know the quality or type of a new user upon arrival, and may need to learn this over time. The platform has various levers at its disposal to address such issues. Such levers can include search engine design and other aspects of the search process, price recommendations, reputation/feedback mechanisms, and the matching algorithm itself in case of centralized matching platforms (e.g., Uber and kidney exchange platforms).

I will stress on the urgent need to understand the tradeoffs between engineering, economic and operational considerations, and hence facilitate the design of better, more efficient marketplaces in matching and related contexts. Along the way, we must constantly draw upon experiences in real markets.